Download (CDT) Research Experience Placement Project Brief

NEXUSS
NEXUSS Centre for Doctoral Training (CDT) Research Experience Placement
Project Brief
Applications close on 2nd June 2017
Lead Supervisor:
Laura Cooper
Email:
[email protected]
University/Research Organisation:
University of Southampton
Department:
Faculty of Engineering and the Environment
Project Title:
Automatic Segmentation of Roots from X-ray
computed tomography images
Total Student Support Costs: £
2000
Based on a minimum of £200/week full time for a minimum of 8 weeks and maximum of 10 weeks
and a £500 Research and Training Support Grant.
Proposed Start Date: 26th June 2017
Proposed End Date: 1st September 2017
Projects should run over the summer vacation and we recommend that projects will have
terminated by the 22th September
Brief Summary – please provide a brief summary (maximum 300 words) of the project.
This should include:
 Project outline;
 Links to staff/School/Centre activity as appropriate;
 Supervisory arrangement;
 How space/equipment/supporting resource demands will be met;
 How the project will enhance the skills of the appointed student;
 Any intellectual property rights concerns that may arise from the work.
Global food availability is a growing concern. Most food comes either directly or indirectly from
plants which, in turn, gather their nutrients from soil via their roots. One of the greatest
challenges in this area is that plants grow in an opaque medium. This can be overcome thanks to
advances in X-ray computer tomography (XCT). These developments have made it possible to
visualise root systems in three dimensions within the soil. However, the resulting 3D image stacks
have similar grey scale levels for both the roots and the surrounding soil. Manual segmentation is
not appropriate for the large number of samples that need to be analysed, in this case 72 scans of
wheat roots. This project requires an automatic segmentation algorithm that is not biased or user
dependent. This algorithm will make it possible to analysis large data sets and investigate the
relationship between plant roots, soil and nutrient sources. This will overcome a major bottleneck
in many current projects undertaken by the Crop Science Engineering Group.
On the first day, the student will be trained by the supervisor, Laura Cooper, and PhD student,
Dan McKay, in using the ImageJ image processing software and previously written codes that have
already been written. For the first two weeks the student will meet informally with the supervisor
daily to receive further training and guidance as they become familiar with the software, codes
and image stacks. After this, supervision meetings will be arranged 1-2 times a week. At the end of
the project the student will have the opportunity to present the methods developed and any
initial results to the Porous Media Group and to train members of the Rooty Team to use the
codes.
The student will be have access to the high performance computers in the μ-VIS computing lab,
which they will be required to apply for, in order to be able to work with the large data sets.
ImageJ is an open source image processing software with online documentation.
This project will give the student the opportunity of working with large data sets and high
performance computers as well as being involved in an active research project investigating rootsoil interaction. The student will enhance their image processing, computer coding, software
engineering and data management skills.
Any codes produced within this project will be made openly available via github.com.
Please give an indicative timescale for the student’s work over the length of the project:
(maximum 150 words).
This should include:
 The broad tasks the student will undertake;
 An indicative timescale for these tasks.






Learn to use ImageJ with the macro scripting language, experiment with the previously
written codes and methods and become familiar with root-soil XCT image stacks, 2 weeks
(including training from supervisor)
Improve upon previous codes or write new codes to automatically segment roots, 2
weeks
Optimise codes for robustness and speed and validate against previously manually
segmented stacks, 2 weeks
Use codes to segment all image stacks, 4-6 weeks (to be run in background with minimal
user input)
Data management, ensuring code is usable by other group members, including writing
documentation and ensuring file names are consistent, 2 weeks
Investigate suitability of code for segmenting other image stacks, 1-2 weeks
Proposed procedure for appointing students, including selection criteria:
Please identify specific criteria that should be considered for the selection of placement students e.g. specific
quantitative skills that may be required, subject knowledge etc. If a student has been pre-selected, or the
research area has been led by the student, please provide the student’s contact details, and a summary of
their suitability for the NEXUSS CDT REP programme.
The appointed student will need basic coding skills and an understanding of statistics and
matrices. It would be an advantage if the student had experience with software engineering,
image processing or signal processing and a background in computer science.